INEOS Köln achieved over 90% Plant Utilization through the use of an AIMMS optimization model to aid decision-making related to the steam cracking process.

After implementing the AIMMS-based DISPO system, INEOS Köln became one of the most effective petrochemical sites for the company.

Bottlenecks are now foreseen at an early stage, allowing INEOS to take appropriate preventive actions, and to avoid, or at least minimize, losses in utilization.

Business Issues

Since the early 1970s, INEOS has used the DISPO production simulation tool to inform production planners of possible constraint violations such as tank overflows, feedstock or product shortages, and necessary TARs of production units. In 1999, due to technology changes and new requirements in the business, Ineos Köln started looking for a new technology that could meet their complex needs and future developments.

The Challenge

INEOS Köln’s first challenge was to capture its highly complex and integrated production processes, involving many different processes for chemicals production with over 160 raw materials and end products, and using over 25 production units such as steam crackers, polyethylene plants, and acrylonitrile plants. The second challenge was to compute the day-to-day production and stock levels, and transport needs, for the coming two years in a reasonable amount of time.

The Solution Part 1: Production simulation

After an initial introduction to AIMMS, a pilot study was conducted to simulate part of the production process, and this convinced INEOS that AIMMS was the way forward. The intuitive formulation in AIMMS allowed INEOS to model their processes quickly, resulting in a ready-to-use DISPO application within a few months.

In order to speed up repetitive time-based simulations, AIMMS has been extended to generate assembly code, allowing INEOS to simulate within a few minutes the complete production over the coming years.

Since 2001, the AIMMS simulation model has played a central role as the ‘Number Cruncher’. Its link to a central SQL database server ensures that new inputs are retrieved and the simulation results are forwarded for action. This capability is enhanced with Insight (an Intranet Reporting tool) designed around the SQL database), and several SAP connections. With the combination of AIMMS, Insight, SQL Server, and SAP, a complete and transparent overview of the past, actual, and future production rates and stock levels is available. As a result, anyone within the entire INEOS organization can access all the information related to production at INEOS Köln.

The Solution Part 2: Process optimization

The DISPO system is currently being extended so that it combines simulation and mathematical programming techniques in order to support steam cracker optimization.

Steam cracking (the manufacturing of olefins from large hydrocarbon molecules) is a core part of INEOS Köln processes. Several input fractions (e.g. naphta, ethane) are fed into the steam cracking furnaces. The outputs of the process are referred to as cracked gas components. A key performance issue is to maximize the material flow through the furnaces under several possible circumstances.

In a joint project between INEOS Köln and AIMMS, an optimization model has been developed to aid decision-making related to the steam cracking process. The relations between the inputs of furnaces and their outputs are described by furnace models, which account for the feed stock characteristics as well as for several furnace parameters (e.g. type, temperature). The dedicated solvers provided by the AIMMS system can solve this optimization model, enabling decisions that are simultaneously optimal with respect to several input factors.

The steam cracking optimization model currently runs as an effective stand-alone AIMMS application. Eventually, it will replace the approach in the DISPO production planning system in which the steam cracking process was modelled sequentially, based on step-by-step computations.

The Benefits

After implementing the AIMMS-based DISPO system, INEOS Köln has become one of the most effective petrochemical sites for the company. Bottlenecks are foreseen at an early stage, allowing INEOS to take appropriate preventive actions, and enabling INEOS to avoid, or at least minimize, losses in utilization. Since everyone can access the production information, anyone can be informed on a need-to-know basis about expected bottlenecks. In this way, the appropriate people can act at the right time on their responsibilities and take appropriate measures.

Furthermore, the steam cracker optimization model brings important improvements on process support compared with the previous computation-based approach.

How did AIMMS add value?

AIMMS, as a powerful, flexible, yet easy-to-learn modeling environment has been the key factor in the successful and rapid redevelopment of DISPO and the steam cracker optimization model. Its intuitive modeling language has enabled a comprehensible and maintainable representation of the complex underlying reality.

About INEOS Köln

Based in Worringen near Cologne, INEOS Köln is one of the largest petrochemical sites worldwide. Established as a joint venture in 1957 (by the name of “EC Erdölchemie GmbH”), it was acquired by BP in 2001. Later, it has been spun out as Innovene, and it now forms part of INEOS group.

INEOS Köln employs 1,900 people and is a key raw materials supplier to the chemical industry. The main base product is naphtha, a by-product from oil refineries, which INEOS Köln uses to manufacture more than 4 million tonnes of raw materials annually, including ethylene (1,100,000 tonnes), polyethylene (530,000 tonnes), ethylene oxide (250,000 tonnes) and butadiene (250,000 tonnes). These are used as starting products for the manufacture of plastics, rubber and fibres, and are also found in solvents, coatings, fertilizers, and crop protection agents.

View this case study as a PDF.